From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <e@80x24.org>
X-Spam-Checker-Version: SpamAssassin 3.4.2 (2018-09-13) on dcvr.yhbt.net
X-Spam-Level: 
X-Spam-ASN:  
X-Spam-Status: No, score=-4.0 required=3.0 tests=ALL_TRUSTED,BAYES_00
	shortcircuit=no autolearn=ham autolearn_force=no version=3.4.2
Received: from localhost (dcvr.yhbt.net [127.0.0.1])
	by dcvr.yhbt.net (Postfix) with ESMTP id 0C4E01F8C8
	for <meta@public-inbox.org>; Mon,  4 Oct 2021 06:10:22 +0000 (UTC)
From: Eric Wong <e@80x24.org>
To: meta@public-inbox.org
Subject: [PATCH] hl_mod: don't memoize highlight::codeGenerator objects
Date: Mon,  4 Oct 2021 06:10:21 +0000
Message-Id: <20211004061021.5825-1-e@80x24.org>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
List-Id: <meta.public-inbox.org>

Making them immortal doesn't seem worth it, since doing immortal
allocations after process startup leads to fragmentation.  While
the allocations made by highlight are small, those small
allocations can break up contiguous regions and prevent
consolidation by the malloc.

Since instantiating code generators doesn't seem too expensive,
just use and delete them ASAP.
---
 lib/PublicInbox/HlMod.pm | 38 +++++++++++++++-----------------------
 t/hl_mod.t               |  3 +--
 2 files changed, 16 insertions(+), 25 deletions(-)

diff --git a/lib/PublicInbox/HlMod.pm b/lib/PublicInbox/HlMod.pm
index 9016db3a..f42ece80 100644
--- a/lib/PublicInbox/HlMod.pm
+++ b/lib/PublicInbox/HlMod.pm
@@ -14,7 +14,7 @@
 # wrapper for SWIG-generated highlight.pm bindings
 package PublicInbox::HlMod;
 use strict;
-use warnings;
+use v5.10.1;
 use highlight; # SWIG-generated stuff
 use PublicInbox::Hval qw(src_escape ascii_html);
 my $hl;
@@ -54,8 +54,7 @@ sub _parse_filetypes ($) {
 	(\%ext2lang, \@shebang);
 }
 
-# We only need one instance, so we don't need to do
-# highlight::CodeGenerator::deleteInstance
+# We only need one instance
 sub new {
 	my ($class) = @_;
 	$hl ||= do {
@@ -95,33 +94,26 @@ sub do_hl {
 sub do_hl_lang {
 	my ($self, $str, $lang) = @_;
 
-	my $dir = $self->{-dir};
 	my $langpath;
-
 	if (defined $lang) {
-		$langpath = $dir->getLangPath("$lang.lang") or return;
-		$lang = undef unless -f $langpath
-	}
-	unless (defined $lang) {
-		$lang = _shebang2lang($self, $str) or return;
-		$langpath = $dir->getLangPath("$lang.lang") or return;
-		return unless -f $langpath
+		$langpath = $self->{-dir}->getLangPath("$lang.lang") or return;
+		undef $lang unless -f $langpath;
 	}
-	my $gen = $self->{$langpath} ||= do {
-		my $g = highlight::CodeGenerator::getInstance($highlight::HTML);
-		$g->setFragmentCode(1); # generate html fragment
+	$lang //= _shebang2lang($self, $str) // return;
+	$langpath = $self->{-dir}->getLangPath("$lang.lang") or return;
+	return unless -f $langpath;
 
-		# whatever theme works
-		my $themepath = $dir->getThemePath('print.theme');
-		$g->initTheme($themepath);
-		$g->loadLanguage($langpath);
-		$g->setEncoding('utf-8');
-		$g;
-	};
+	my $g = highlight::CodeGenerator::getInstance($highlight::HTML);
+	$g->setFragmentCode(1); # generate html fragment
 
+	# whatever theme works
+	$g->initTheme($self->{-dir}->getThemePath('print.theme'));
+	$g->loadLanguage($langpath);
+	$g->setEncoding('utf-8');
 	# we assume $$str is valid UTF-8, but the SWIG binding doesn't
 	# know that, so ensure it's marked as UTF-8 even if it isnt...
-	my $out = $gen->generateString($$str);
+	my $out = $g->generateString($$str);
+	highlight::CodeGenerator::deleteInstance($g);
 	utf8::decode($out);
 	src_escape($out);
 	\$out;
diff --git a/t/hl_mod.t b/t/hl_mod.t
index 96878d51..a88f6c03 100644
--- a/t/hl_mod.t
+++ b/t/hl_mod.t
@@ -3,8 +3,7 @@
 # License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
 use strict; use v5.10.1; use PublicInbox::TestCommon; use IO::Handle; # ->autoflush
 use Fcntl qw(:seek);
-eval { require highlight } or
-	plan skip_all => "failed to load highlight.pm for $0";
+require_mods 'highlight';
 use_ok 'PublicInbox::HlMod';
 my $hls = PublicInbox::HlMod->new;
 ok($hls, 'initialized OK');